Discovery of novel Plasmodium falciparum PfDHFR-TS inhibitors from ConMedNP natural compounds: a multi-computational approach.

Abstract

The rise of drug-resistant Plasmodium falciparum necessitates novel antimalarial therapies. Leveraging the ConMedNP database, which includes over 3119 natural compounds from Central and West African medicinal plants, this study targets Plasmodium falciparum dihydrofolate reductase-thymidylate synthase (PfDHFR-TS), a vital enzyme for parasite survival. Molecular docking of 2754 compounds revealed a mean binding affinity of $-$  8.8032 kcal/mol (SD = 1.4 kcal/mol, median = $-$  8.9 kcal/mol), with 75% outperforming artemether's reference affinity ( $-$  8.0 kcal/mol). A Random Forest-based RaMQSAR model, trained on the docking data, achieved a test $R^2$ of 0.8321 (RMSE: 0.5294 kcal/mol) and reliable cross-validation (mean $R^2$ = 0.8461, SD = 0.0460). Validation against 19 known antimalarials showed predicted affinities from $-$  7.0 to $-$  10.5 kcal/mol, consistent with docking results. Top performers included RDC0118 ( $-$  13.5 kcal/mol), RDC0119 ( $-$  13.4 kcal/mol), and CA0001 ( $-$  13.0 kcal/mol), all surpassing artemether. ADMET profiling indicated CA0001 and artemether as safer candidates (non-hepatotoxic, low environmental impact), while RDC0118 and RDC0119 exhibited potential mutagenicity and hepatotoxicity risks. MD simulations confirmed structural stability for both, with CA0001 showing compaction and transient H-bonds (0-3). DFT analysis highlighted CA0001's reactivity as a soft electrophile, contrasting with artemether's higher reactivity. This comprehensive approach integrating docking, QSAR, DFT, and MD positions CA0001 as a promising PfDHFR-TS inhibitor alongside artemether, with ConMedNP and predictive models guiding future experimental validation.